Long‐term premonitory seismicity patterns in Tibet and the Himalayas

Abstract

An attempt is made to identify seismicity patterns precursory to great earthquakes in most of Tibet as well as the central and eastern Himalayas. The region has considerable tectonic homogeneity and encompasses parts of China, India, Nepal, Bhutan, Bangladesh, and Burma. Two seismicity patterns previously described were used: (1) pattern Σ is a peak in the sum of earthquake energies raised to the power of about 2/3, taken over a sliding time window and within a magnitude range less than that of events we are trying to predict; and (2) pattern S (swarms) consists of the spatial clustering of earthquakes during a time interval when the seismicity is above average. Within the test region, distinct peaks in pattern Σ have occurred twice during the 78‐year‐long test period: in 1948–49, prior to the great 1950 Assam‐Tibet earthquake (M = 8.6), and in 1976. Peaks in pattern S have occurred three times; in 1932–1933, prior to the great 1934 Bihar‐Nepal earthquake (M = 8.3), in 1946, and in 1978. The 1934 and 1950 earthquakes were the only events in the region that exceeded M = 8.0 during the test period. On the basis of experience here and elsewhere, the current peaks in both Σ and S suggest the likelihood of an M = 8.0 event within 6 years or an M = 8.5 event within 14 years. Such a prognostication should be viewed more as an experimental long‐term enhancement of the probability that a large earthquake will occur than as an actual prediction, in view of the exceedingly large area encompassed and the very lengthy time window. Furthermore, the chances of a randomly occurring event as large as M = 8.0 in the region are perhaps 21% within the next 6 years, and the present state of the art is such that we can place only limited confidence in such forecasts. The primary impact of the study, in our opinion, should be to stimulate the search for medium‐ and short‐term precursors in the region and to search for similar long‐term precursors elsewhere.